Rust: Add some flow source models

rust/ql/lib/codeql/rust/Concepts.qll

@@ -6,6 +6,7 @@
 
 private import codeql.rust.dataflow.DataFlow
 private import codeql.threatmodels.ThreatModels
+private import codeql.rust.Frameworks
 
 /**
  * A data flow source for a specific threat-model.
@@ -46,6 +47,63 @@ class ActiveThreatModelSource extends ThreatModelSource {
   ActiveThreatModelSource() { currentThreatModel(this.getThreatModel()) }
 }
 
+/**
+ * A data flow source corresponding to the program's command line arguments or path.
+ */
+final class CommandLineArgsSource = CommandLineArgsSource::Range;
+
+/**
+ * Provides a class for modeling new sources for the program's command line arguments or path.
+ */
+module CommandLineArgsSource {
+  /**
+   * A data flow source corresponding to the program's command line arguments or path.
+   */
+  abstract class Range extends ThreatModelSource::Range {
+    override string getThreatModel() { result = "commandargs" }
+
+    override string getSourceType() { result = "CommandLineArgs" }
+  }
+}
+
+/**
+ * A data flow source corresponding to the program's environment.
+ */
+final class EnvironmentSource = EnvironmentSource::Range;
+
+/**
+ * Provides a class for modeling new sources for the program's environment.
+ */
+module EnvironmentSource {
+  /**
+   * A data flow source corresponding to the program's environment.
+   */
+  abstract class Range extends ThreatModelSource::Range {
+    override string getThreatModel() { result = "environment" }
+
+    override string getSourceType() { result = "EnvironmentSource" }
+  }
+}
+
+/**
+ * A data flow source for remote (network) data.
+ */
+final class RemoteSource = RemoteSource::Range;
+
+/**
+ * Provides a class for modeling new sources of remote (network) data.
+ */
+module RemoteSource {
+  /**
+   * A data flow source for remote (network) data.
+   */
+  abstract class Range extends ThreatModelSource::Range {
+    override string getThreatModel() { result = "remote" }
+
+    override string getSourceType() { result = "RemoteSource" }
+  }
+}
+
 /**
  * A data-flow node that constructs a SQL statement.
  *

rust/ql/lib/codeql/rust/Frameworks.qll

@@ -0,0 +1,6 @@
+/**
+ * This file imports all models of frameworks and libraries.
+ */
+
+private import codeql.rust.frameworks.Reqwest
+private import codeql.rust.frameworks.stdlib.Env

rust/ql/lib/codeql/rust/frameworks/Reqwest.qll

@@ -0,0 +1,19 @@
+/**
+ * Provides modeling for the `reqwest` library.
+ */
+
+private import rust
+private import codeql.rust.Concepts
+
+/**
+ * A call to `reqwest::get` or `reqwest::blocking::get`.
+ */
+private class ReqwestGet extends RemoteSource::Range {
+  ReqwestGet() {
+    exists(CallExpr ce |
+      this.asExpr().getExpr() = ce and
+      ce.getExpr().(PathExpr).getPath().getResolvedCrateOrigin().matches("%reqwest") and
+      ce.getExpr().(PathExpr).getPath().getResolvedPath() = ["crate::get", "crate::blocking::get"]
+    )
+  }
+}

rust/ql/lib/codeql/rust/frameworks/stdlib/Env.qll

@@ -0,0 +1,36 @@
+/**
+ * Provides modeling for the `std::env` library.
+ */
+
+private import rust
+private import codeql.rust.Concepts
+
+/**
+ * A call to `std::env::args` or `std::env::args_os`.
+ */
+private class StdEnvArgs extends CommandLineArgsSource::Range {
+  StdEnvArgs() {
+    this.asExpr().getExpr().(CallExpr).getExpr().(PathExpr).getPath().getResolvedPath() =
+      ["crate::env::args", "crate::env::args_os"]
+  }
+}
+
+/**
+ * A call to `std::env::current_dir`, `std::env::current_exe` or `std::env::home_dir`.
+ */
+private class StdEnvDir extends CommandLineArgsSource::Range {
+  StdEnvDir() {
+    this.asExpr().getExpr().(CallExpr).getExpr().(PathExpr).getPath().getResolvedPath() =
+      ["crate::env::current_dir", "crate::env::current_exe", "crate::env::home_dir"]
+  }
+}
+
+/**
+ * A call to `std::env::var`, `std::env::var_os`, `std::env::vars` or `std::env::vars_os`.
+ */
+private class StdEnvVar extends EnvironmentSource::Range {
+  StdEnvVar() {
+    this.asExpr().getExpr().(CallExpr).getExpr().(PathExpr).getPath().getResolvedPath() =
+      ["crate::env::var", "crate::env::var_os", "crate::env::vars", "crate::env::vars_os"]
+  }
+}

rust/ql/src/queries/summary/SummaryStats.ql

@@ -7,6 +7,7 @@
  */
 
 import rust
+import codeql.rust.Concepts
 import codeql.rust.Diagnostics
 import Stats
 
@@ -43,4 +44,8 @@ where
   key = "Macro calls - resolved" and value = count(MacroCall mc | mc.hasExpanded())
   or
   key = "Macro calls - unresolved" and value = count(MacroCall mc | not mc.hasExpanded())
+  or
+  key = "Taint sources - total" and value = count(ThreatModelSource s)
+  or
+  key = "Taint sources - active" and value = count(ActiveThreatModelSource s)
 select key, value order by key

rust/ql/src/queries/summary/TaintSources.ql

@@ -0,0 +1,18 @@
+/**
+ * @name Taint Sources
+ * @description List all sources of untrusted input that have been idenfitied
+ *              in the database.
+ * @kind problem
+ * @problem.severity info
+ * @id rust/summary/taint-sources
+ * @tags summary
+ */
+
+import rust
+import codeql.rust.Concepts
+
+from ThreatModelSource s, string defaultString
+where
+  if s instanceof ActiveThreatModelSource then defaultString = " (DEFAULT)" else defaultString = ""
+select s,
+  "Flow source '" + s.getSourceType() + "' of type " + s.getThreatModel() + defaultString + "."

rust/ql/test/library-tests/dataflow/sources/InlineFlow.ql

@@ -0,0 +1,21 @@
+import rust
+import codeql.rust.dataflow.DataFlow
+import codeql.rust.Concepts
+import utils.InlineFlowTest
+
+/**
+ * Configuration for flow from any threat model source to an argument of the function `sink`.
+ */
+module MyFlowConfig implements DataFlow::ConfigSig {
+  predicate isSource(DataFlow::Node source) { source instanceof ThreatModelSource }
+
+  predicate isSink(DataFlow::Node sink) {
+    any(CallExpr call | call.getExpr().(PathExpr).getPath().getResolvedPath() = "crate::test::sink")
+        .getArgList()
+        .getAnArg() = sink.asExpr().getExpr()
+  }
+}
+
+module MyFlowTest = TaintFlowTest<MyFlowConfig>;
+
+import MyFlowTest

rust/ql/test/library-tests/dataflow/sources/TaintSources.expected

@@ -0,0 +1,17 @@
+| test.rs:8:10:8:30 | ...::var(...) | Flow source 'EnvironmentSource' of type environment. |
+| test.rs:9:10:9:33 | ...::var_os(...) | Flow source 'EnvironmentSource' of type environment. |
+| test.rs:11:16:11:36 | ...::var(...) | Flow source 'EnvironmentSource' of type environment. |
+| test.rs:12:16:12:39 | ...::var_os(...) | Flow source 'EnvironmentSource' of type environment. |
+| test.rs:17:25:17:40 | ...::vars(...) | Flow source 'EnvironmentSource' of type environment. |
+| test.rs:22:25:22:43 | ...::vars_os(...) | Flow source 'EnvironmentSource' of type environment. |
+| test.rs:29:29:29:44 | ...::args(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:32:16:32:31 | ...::args(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:33:16:33:34 | ...::args_os(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:40:16:40:31 | ...::args(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:44:16:44:34 | ...::args_os(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:50:15:50:37 | ...::current_dir(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:51:15:51:37 | ...::current_exe(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:52:16:52:35 | ...::home_dir(...) | Flow source 'CommandLineArgs' of type commandargs. |
+| test.rs:60:26:60:70 | ...::get(...) | Flow source 'RemoteSource' of type remote (DEFAULT). |
+| test.rs:63:26:63:70 | ...::get(...) | Flow source 'RemoteSource' of type remote (DEFAULT). |
+| test.rs:66:26:66:60 | ...::get(...) | Flow source 'RemoteSource' of type remote (DEFAULT). |

rust/ql/test/library-tests/dataflow/sources/TaintSources.qlref

@@ -0,0 +1,2 @@
+query: queries/summary/TaintSources.ql
+postprocess: utils/InlineExpectationsTestQuery.ql

rust/ql/test/library-tests/dataflow/sources/options.yml

@@ -0,0 +1,3 @@
+qltest_cargo_check: true
+qltest_dependencies:
+    - reqwest = { version = "0.12.9", features = ["blocking"] }

rust/ql/test/library-tests/dataflow/sources/reqwest.rs

@@ -0,0 +1,36 @@
+
+// --- stubs for the "reqwest" library ---
+
+/*
+ --- we don't seem to have a way to use this, hence we currently test against the real reqwest library
+#[derive(Debug)]
+pub struct Error { }
+
+pub mod blocking {
+    pub struct Response { }
+    impl Response {
+        pub fn text(self) -> Result<String, super::Error> {
+            Ok("".to_string())
+        }
+    }
+
+    pub fn get<T>(url: T) -> Result<Response, super::Error> {
+        let _url = url;
+
+        Ok(Response {})
+    }
+}
+
+pub struct Response { }
+impl Response {
+    pub async fn text(self) -> Result<String, Error> {
+        Ok("".to_string())
+    }
+}
+
+pub async fn get<T>(url: T) -> Result<Response, Error> {
+    let _url = url;
+
+    Ok(Response {})
+}
+*/

rust/ql/test/library-tests/dataflow/sources/test.rs

@@ -0,0 +1,70 @@
+#![allow(deprecated)]
+
+fn sink<T>(_: T) { }
+
+// --- tests ---
+
+fn test_env_vars() {
+    sink(std::env::var("HOME")); // $ Alert[rust/summary/taint-sources] hasTaintFlow
+    sink(std::env::var_os("PATH")); // $ Alert[rust/summary/taint-sources] hasTaintFlow
+
+    let var1 = std::env::var("HOME").expect("HOME not set"); // $ Alert[rust/summary/taint-sources]
+    let var2 = std::env::var_os("PATH").unwrap(); // $ Alert[rust/summary/taint-sources]
+
+    sink(var1); // $ MISSING: hasTaintFlow
+    sink(var2); // $ MISSING: hasTaintFlow
+
+    for (key, value) in std::env::vars() { // $ Alert[rust/summary/taint-sources]
+        sink(key); // $ MISSING: hasTaintFlow
+        sink(value); // $ MISSING: hasTaintFlow
+    }
+
+    for (key, value) in std::env::vars_os() { // $ Alert[rust/summary/taint-sources]
+        sink(key); // $ MISSING: hasTaintFlow
+        sink(value); // $ MISSING: hasTaintFlow
+    }
+}
+
+fn test_env_args() {
+    let args: Vec<String> = std::env::args().collect(); // $ Alert[rust/summary/taint-sources]
+    let my_path = &args[0];
+    let arg1 = &args[1];
+    let arg2 = std::env::args().nth(2).unwrap(); // $ Alert[rust/summary/taint-sources]
+    let arg3 = std::env::args_os().nth(3).unwrap(); // $ Alert[rust/summary/taint-sources]
+
+    sink(my_path); // $ MISSING: hasTaintFlow
+    sink(arg1); // $ MISSING: hasTaintFlow
+    sink(arg2); // $ MISSING: hasTaintFlow
+    sink(arg3); // $ MISSING: hasTaintFlow
+
+    for arg in std::env::args() { // $ Alert[rust/summary/taint-sources]
+        sink(arg); // $ MISSING: hasTaintFlow
+    }
+
+    for arg in std::env::args_os() { // $ Alert[rust/summary/taint-sources]
+        sink(arg); // $ MISSING: hasTaintFlow
+    }
+}
+
+fn test_env_dirs() {
+    let dir = std::env::current_dir().expect("FAILED"); // $ Alert[rust/summary/taint-sources]
+    let exe = std::env::current_exe().expect("FAILED"); // $ Alert[rust/summary/taint-sources]
+    let home = std::env::home_dir().expect("FAILED"); // $ Alert[rust/summary/taint-sources]
+
+    sink(dir); // $ MISSING: hasTaintFlow
+    sink(exe); // $ MISSING: hasTaintFlow
+    sink(home); // $ MISSING: hasTaintFlow
+}
+
+async fn test_reqwest() -> Result<(), reqwest::Error> {
+    let remote_string1 = reqwest::blocking::get("http://example.com/")?.text()?; // $ Alert[rust/summary/taint-sources]
+    sink(remote_string1); // $ MISSING: hasTaintFlow
+
+    let remote_string2 = reqwest::blocking::get("http://example.com/").unwrap().text().unwrap(); // $ Alert[rust/summary/taint-sources]
+    sink(remote_string2); // $ MISSING: hasTaintFlow
+
+    let remote_string3 = reqwest::get("http://example.com/").await?.text().await?; // $ Alert[rust/summary/taint-sources]
+    sink(remote_string3); // $ MISSING: hasTaintFlow
+
+    Ok(())
+}

rust/ql/test/query-tests/diagnostics/SummaryStats.expected

@@ -13,3 +13,5 @@
 | Macro calls - resolved | 8 |
 | Macro calls - total | 9 |
 | Macro calls - unresolved | 1 |
+| Taint sources - active | 0 |
+| Taint sources - total | 0 |